PROJECT SUMMARY Quantitative tracer images from combined PET-CT and PET-MR are increasingly being utilized for radiological decision-making and clinical trials. CT and MR based data corrections for PET photon attenuation, typically the largest impact on tracer quantification, can lead to increased radiation dose to the subject or require lengthy additional dedicated MR scans, respectively. Artificial intelligence algorithms that use the PET signal originating from the subject alone to estimate attenuation have shown promising early results, but, can have reduced performance when applied to exams different from the training data. Reconstruction methods that jointly estimate both the measured attenuation and tracer contrast from the subject PET signal represent a viable alternative, but, to date, have failed to match the quantification of CT approaches. Consequently, no single approach for PET attenuation correction that consistently produces high quantification and does not compromise patient safety or throughput currently exists. This is an important problem, since research and clinical findings must balance these practical issues with data quality. The overall objective of this proposal is to develop and characterize a high performance attenuation correction scheme that utilizes both the PET signal from the subject and a variable activity external source. The central hypothesis is that the proposed reconstruction method will have significantly improved performance over reconstruction algorithms using PET signal originating from the subject alone, independent of the imaging study. The two specific aims include: 1) developing and validating a supplemental transmission-based algorithm for correcting PET images for photon attenuation during commercial PET imaging and 2) prototyping and characterizing a device capable of dynamically varying external source activity. The result of Aim 1 will be a reconstruction algorithm that is optimized to produce quantitative PET data for some of the most common clinical PET studies. Under Aim 2, a prototype device that can repeatability vary the external source activity in order to maximize tracer quantification while minimizing the unavoidable degradation to patient tracer image noise, caused by the introduction of any external source, will be produced. The innovation is an attenuation correction strategy that greatly mitigates the limitations of previous joint reconstruction methods to deliver quantitative PET diagnostics that are expected to match those of silver standard CT approaches. This is significant because the number of patients receiving brain and cardiac PET-CT scans, exams the proposed method is expected to benefit, is steadily increasing. For PET-MR, the method may improve tracer quantification where MR has limited performance and increase patient throughput by eliminating the need for non-diagnostic attenuation-only MR acquisitions. Thus, the proposed strategy has great potential to significantly improve radiological decision- making and clinical trial findings that rely on quantitative PET uptake measurements.